Gas Flow at the Ultra-nanoscale: Universal Predictive Model and

Publication Date (Web): September 5, 2018. Copyright © 2018 American Chemical Society. *E-mail: [email protected]. Cite this:ACS Appl. M...
0 downloads 0 Views 876KB Size
Subscriber access provided by University of South Dakota

Energy, Environmental, and Catalysis Applications

Gas Flow at the Ultra-nanoscale: Universal Predictive Model and Validation in Nanochannels of Ångstrom-Level Resolution Giovanni Scorrano, Giacomo Bruno, Nicola Di Trani, Mauro Ferrari, Alberto Pimpinelli, and Alessandro Grattoni ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.8b11455 • Publication Date (Web): 05 Sep 2018 Downloaded from http://pubs.acs.org on September 10, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

Table of Content image 50x31mm (300 x 300 DPI)

ACS Paragon Plus Environment

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Gas Flow at the Ultra-nanoscale: Universal Predictive Model and Validation in Nanochannels of Ångstrom-Level Resolution Giovanni Scorrano,1,2 Giacomo Bruno,1 Nicola Di Trani,1 Mauro Ferrari,1 Alberto Pimpinelli,1,3 and Alessandro Grattoni1,4,5* 1

Department of Nanomedicine, Houston Methodist Research Institute, Houston, Texas,77030, USA

2

Department of Material Science and Nanoengineering, Rice University, Houston, Texas, 77005, USA

3

Smalley-Curl Institute, Rice University, Houston, Texas, 77005, USA.

4

Department of Surgery, Houston Methodist Hospital, Houston, Texas, 77030, USA.

5

Department of Radiation Oncology, Houston Methodist Hospital, Houston, Texas, 77030, USA.

*corresponding author email: [email protected]

Keywords: nanochannels, convective gas flow, Knudsen regime, nanofluidic membrane

Abstract Gas transport across nanoscale pores is determinant in molecular exchange in living organisms as well as in a broad spectrum of technologies. Here we report an unprecedented theoretical and experimental analysis of gas transport in a consistent set of confining nanochannels ranging in size from the ultra-nano- to the sub-microscale. A generally applicable theoretical approach quantitatively predicting confined gas flow in the Knudsen and transition regime was developed. Unlike current theories, specifically designed for very simple channel geometries, our approach can be applied to virtually all geometries, for which the probability distribution of path lengths for particle-interface collisions can be computed, either analytically or by numerical simulations. To generate a much needed benchmark experimental model, we manufactured extremely reproducible membranes with two-dimensional nanochannels. Channel sizes ranged from 2.5 to 250 nm, and Å-level of size control and interface tolerances were achieved using leading-edge

ACS Paragon Plus Environment

Page 2 of 22

Page 3 of 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

nanofabrication techniques. We then measured gas flow in the Knudsen number range from 0.2 to 20. Excellent agreement between theoretical predictions and experimental data was found, demonstrating the validity and potential of our approach.

1. Introduction Numerous medical and engineering applications rely on gas flow in nanopores where the collisions of fluid molecules with confining walls dominate the characteristics of gas transport. A highly relevant example is represented by natural gas extraction from nanoporous shale rock1. Here, understanding nano-confined transport becomes extremely important for designing extraction techniques as well for predicting trends of gas production. Other examples are represented by gas filtration in chemical plants2,3,4 or the development and characterization of high efficiency membranes for applications such as molecular sieving5,6 or drug7,8,9 and cell delivery10,11 among others. These involve gas molecules traveling across tight spaces where the interactions with the walls are predominant as compared to the interactions among gas molecules. Furthermore, in the context of heat transport in nanoscale systems, the mean free path of a confined phonon gas is an essential ingredient in models of thermal conductivity12. Ideally, one can identify two limiting regimes, characterized by distinct length scales: i) A viscous regime, in which the thermal mean free path λT ≪ LC, where LC is the characteristic length scale

of the confining nanostructure; ii) A rarefied (free molecular flow) regime, in which λT ≫ LC.

The latter was first investigated by Knudsen13. Accordingly, the Knudsen number Kn is defined as the ratio Kn = λT / LC. On a finer scale, three regions can be identified: i) when Kn < 0.1, the gas is treated as a continuum and the Navier-Stokes equation with slip boundary condition

ACS Paragon Plus Environment

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

governs gas transport14; ii) when Kn > 10, gas collisions with the confining walls dominate; this free molecular flow regime is also known as Knudsen diffusion; iii) an intermediate region, 0.1 < Kn < 10, called transition regime, has been experimentally and theoretically explored in the literature, primarily for gases at low pressure in microfluidic systems15. Pressure-driven gas transport in this latter regime is characterized by the coexistence of diffusive and convective (advective) transport. As is the case for concentration driven diffusive transport16,17, the formal description of gas flow in the transition regime in nanofluidic channels remains an open question. Surprisingly, free molecular flow is also debated: Knudsen’s results rigorously apply only to gas flow inside cylindrical channels, and although his work is more than a century old, its generalization to other geometries remains controversial18. Experimental studies of nanoconfined gas flow are challenging, especially because of technological limitations in the fabrication of precise nanofluidic systems with a high number and density of channels and tight dimensional and geometrical tolerances. Most investigations have relied on experimental data collected with a few channels of irregular cross section, approximate size14, and minuscule gas flow outputs19. Experiments adopting polymeric membranes have been limited by intricate pore geometries, poor reproducibility or high channel tortuosity18. A variety of porous materials such as zeolites20, and carbon materials21,22, such as carbon nanotubes (CNT), have received considerable attention due to their confining fluidic properties. However, their intrinsic geometric and structural variability22 limit the investigator’s ability to study the underlying transport phenomena; one can at best know the average properties of these structures, such as their average pore size - thus preventing the study of gas transport as a function of the actual Knudsen number. By leveraging leading-edge, industrial grade microfabrication techniques, we have manufactured highly accurate fluidic structures with tightly reproducible geometry, ranging in size from the ultra-

ACS Paragon Plus Environment

Page 4 of 22

Page 5 of 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

nanoscale to the sub-microscale23. By offering an unmatched control on channel size and geometry, such structures represent an ideal model system for the accurate investigation of the influence of geometry and size on gas transport24 from the Knudsen to the transition regime25.

2. Materials and Methods 2.1 Model derivation The kinetic theory of gases allows to compute the distribution of paths of length l between intermolecular collisions26 in thermal equilibrium, pT(l), whose mean value is the thermal mean free path λT. For a gas of identical spherical molecules of diameter d, where velocities are Maxwell-distributed at temperature T and pressure P, the thermal mean free path reads as 

λ = 





, where kB is the Boltzmann’s constant,  is the gas viscosity and m the molecular

mass. It should be noted that the expectation that pT is exponential: pT(l) = exp(−l /λ  )dl /λ  , is generally not true. For simplicity, we will treat it as exponential in the applications that follow. Consider a gas in a channel of rectangular section h × w and length L. When a pressure difference △P = Pin − Pout is applied between the channel’s ends, a flux F results. Kinetic theory states that the volumetric gas flow is

=

ℎ ∆ 4λ λ  ℎ = ∆   3 

(1)

where m is the molecular mass and D = λ /3 is the gas diffusion coefficient written in terms of the mean thermal velocity c and the average distance between consecutive collisions, λ . The

ACS Paragon Plus Environment

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

latter is the characteristic length that has to be computed as a function of the geometry. At Knudsen numbers Kn ≤ 1, it is expected that λ ≈ λ  in Eq. (1). When Kn >> 1, collisions between gas molecules can be neglected, and particle-wall encounters dominate the gas transport. In such conditions, λ0 ≈ λg, the latter being the average distance between successive collisions of a gas molecule with the channel’s walls. Knudsen computed λg for a cylindrical geometry and showed that λg = 2R. He then computed λg for a long rectangular channel of uniform cross-section h × w, where h, w ≪ L, and found that λg= 2hw/(h+w). For a gas confined between two infinite parallel plates (w → ∞), λg = 2h. Stops27 computed the effective mean free path in the transition regime for two infinite plates at distance h. By considering that a number of molecular paths are prematurely terminated by the boundaries, so that the effective mean free path λ0 is less than the thermal mean free path λT, he computed an explicit expression, given in Eq. (8) in Stops27. We re-derive Stops’ λ0 and prove that his result is exact. Stops’ formulation is ad hoc for the infinite parallel plates geometry; it did not consider that the geometric mean free path λg can be defined as the average of an appropriate distribution of path lengths, in the same way as λT can. Such distributions, that we call “geometric distribution of path lengths” (pg(l)), are well known in the theory of the Lorentz gas,28, 29 as well as in classical billiard dynamics. For the latter, pg(l) has been recently published by Holmin et al.30 (see the full expression for a 3D box in Supporting Information Eq. S1-2). Having defined both λT and λg as the mean values of pT(l) and pg(l), respectively, we can leverage the statistical independence of particle-wall and particle-particle collisions. Since pT(l) is the probability that a molecule collides with another molecule after traveling a distance l, the total probability that a molecule travels a distance l without collisions with other molecules is *

)

equal to 1 − # $ %&'(& = #* $ %&'(&. Similarly, the probability that a molecule travels a

ACS Paragon Plus Environment

Page 6 of 22

Page 7 of 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

)

distance l without collisions with walls reads #* $+ %&'(&. Statistical independence guarantees that the probability of a molecule colliding with a wall but not with other molecules after )

traveling a distance l is $+ %,' #* $ %&'(&. Interchanging the subscripts “T” and “g” yields the probability of a molecule colliding with another molecule, but not with a wall: )

$ %,' #* $+ %&'(&. 2.2 Experimental setup To test our model against experimental data, we developed and micro-fabricated several sets of nanoslit membranes, all identical except for the height h of the nanoslits. The membranes (Fig. 1) were fabricated through a sacrificial layer technique. By precisely tailoring the thickness of a tungsten film deposited by physical vapor deposition with angstrom-level resolution, slit channels with nominal heights of, respectively, 2.5, 3.5, 20, 50, 200, and 250 nm, parallel to the membrane surface (Fig. 1A) were obtained. During the fabrication process the tungsten layer represented the “space keeper” for the nanochannels. Details of the membrane fabrication are available elsewhere23,24 and in Supporting Information. Scanning and transmission electron microscopy image analysis showed deviations from the nominal height of less than 5 Å (Fig. 1BE). Each membrane is composed of exactly 340,252 nanochannels regularly and densely organized in rectangular arrays and connected to the membrane inlet and outlet surfaces via arrays of microchannels (Fig. 1A).

ACS Paragon Plus Environment

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 1: Schematic depiction of the membrane structure (A); TEM cross-sectional images of selected membranes (B, C, D, E); see text for details.

This design affords high channel density and mechanical robustness23 and unparalleled dimensional tolerances as compared with other channel geometries. Indeed, even the most refined membranes presenting cylindrical nanopores (e.g. TiO2 or carbon nanotubes) suffer from much poorer dimensional accuracy and manufacturing control, as well as lower scalability. We verified that despite the presence of microchannels, gas flow across our membranes is entirely dominated by the parallel set of nanochannels that represent the dominant fluidic resistance. Finite element simulations (COMSOL Multiphysics) showed that in first analysis, even for the membranes with the largest nanochannels (h = 250 nm), the pressure drop is localized between the inlet and outlet of the nanochannel region (Supporting Information Fig. S1), accounting for more than 95% for the pressure drop across the membranes, when considering entrance and exit effects. Gas flow through silicon membranes (Fig. 1A) possessing rectangular slit-nanochannels (W = 3µm, L = 1µm, and variable h) was measured with a high sensitivity apparatus consisting in a

ACS Paragon Plus Environment

Page 8 of 22

Page 9 of 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

pressure controller (PC3-15PSIG, Alicat) and three mass flow meters (M-500SCCM, M10SCCM, M-0.5SCCM, Alicat) with different sensitivity to ensure a high signal to noise ratio always larger than 13.85 (Fig. 2). Measurements were performed by a custom-made algorithm (Matlab, The MathWorks, Inc.) setting several differential pressures from 1 to 15 psi and recording the gas flow through the silicon membrane in steady state conditions.

Figure 2. Schematics of the high sensitivity apparatus developed for mass flow measurements.

3. Results and Discussion The derivation of probabilities of particle-particle and particle-wall collisions allowed us to determine the probability distribution of path lengths in a confined region of arbitrary geometry

ACS Paragon Plus Environment

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

)

Page 10 of 22

)

$+ %,' #* $ %&'(& - $ %,' #* $+ %&'(&, and the effective mean free path λ0 of a gas confined in a container of arbitrary geometry: )

)

)

λ = . , /$+ %,' . $ %&'(& - $ %,' . $+ %&'(&0 (,. 

*

*

(2)

Eq. (2) is the first important result of this paper. Eq. (2) is completely general, and for it to be useful one has to prescribe pT and pg. For instance, assuming both pT and pg to be exponential, the result λ0-1 = λT-1 +λg-1 is recovered. The calculations of the effective mean free path λ0 in an infinite channel with rectangular cross section or in the parallel planes geometry27 are not trivial. From Eq. (2), λ0 can be obtained through integrations: pT(l) can be found in Paik31 or assumed to be exponential, while pg(l) can be found in Holmin et al.30 and in Supporting Information (Eq. (S1-2)). The resulting λ0 is shown in Fig. 3 (red solid line) for a gas with thermal mean free path λT = 44.5 nm in a box of dimensions w = 3 µm, L = 1 µm.

ACS Paragon Plus Environment

Page 11 of 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

Figure 3. Comparison of mean free path variation for the present model (solid red line), results from Stops (dashed black line), and Knudsen’s model (solid blue line).

Figure 3 also shows λg, as obtained analytically by Holmin et al. for a box of sides h, w and L (solid blue line): λg = 2hwL/(hw + hL + wL). This coincides with Knudsen’s λg = 2hw/(h+w); for an infinite channel, L →∞, and that for infinite parallel planes (w → ∞), λg = 2h. In this latter case, the geometric path distribution pg∞(l) can be found when w, L →∞ in the geometric path distribution of Holmin et al.30 (Supporting Information Eq. (S1-2)):

0, 567 , < ℎ $+) %,' = 2 9: 6?= ;

(3)

*

Alternatively, Eq. (3) can be derived from Lambert’s cosine law, pθ (θ) = 2sinθcosθ, together with the relations pg∞(l) dl = pθ(θ)dθ, l = h/cosθ. From Eq. (3), assuming pT(l)=exp(-l/ λT)/ λT, and performing integrations, yields: ) λ @A @A

= 1 − = - B= − B . & @C = @D (&, λ A

(4)

which coincides with Eq. (8) in Stops27 where β = h/λT . Eq.(4) is also shown in Fig. 3 for comparison (dashed black line). More complex situation can be addressed. Since Eq. (2) is an exact expression, it allows to compute the mean free path of a gas in a channel for any channel shape and wall roughness, provided pg(l) is known. In general, pg(l) has to be computed numerically, e.g. by Monte Carlo simulations which are routinely employed to evaluate the diffusion coefficient in the Knudsen regime inside channels with complicated geometries and/or rough walls. Surface roughness was

ACS Paragon Plus Environment

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 22

shown to lower the diffusion coefficient32. Thus, we propose to generalize Eq. (1) by replacing D with D’ = α0 D, and argue that the roughness-dependent α0 plays a similar role in the definition of the accommodation coefficient of the continuum Navier-Stokes equation with Maxwell boundary conditions, when collisions with the walls are properly accounted for33,34. Arkilic et al have derived an analytical expression for gas flow through a rectangular channel, from the zeroth order solution of the compressible Navier-Stokes equation35. Rewriting their mass flow (Eq. (21) in Arkilic et al.35) as a volumetric flow Q, and introducing the average pressure E = %FG HIJ '/2, yields

L=

ℎM  E 2−O / -6 λ  0 ∆, 12 HIJ O

(5)

where %2 − O'/O is the streamwise momentum accommodation coefficient. At Kn >> 1, Eq. (5) must agree with Eq. (1) when α0 D replaces D; this implies that Defining O = O%ℎ = 0', one can write explicitely:

O%ℎ' =

4O ℎ . 2λ - O %2ℎ − λ '

@P P

=

QRS TS M 9

, so that O = O%ℎ'.

(6)

From λ ~2ℎ as ℎ → 0, one finds that W = 3%2 − O '/%16O ', so that for diffusively reflecting walls, O = 1 implies W ≈ 0.589. We can therefore rewrite Eq. (5) as follows:

L=

ℎM  E 2 − O λ %ℎ'λ  / -3 0 ∆. 12 HIJ O ℎ

(7)

Equation (7) is the second important result of the present work, because only Navier-Stokes equations with geometry-dependent accommodation coefficient are able to reproduce

ACS Paragon Plus Environment

Page 13 of 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

experimentally observed Knudsen minimum in the channel permeability [ =

M \ Q

E

9: ] ∆

^

√`

. We

demonstrate this in Figure 4 where the results of permeability as a function of δ = 1/Kn are shown. Figure 4 displays Eq. (5) in Arkilic et al35 with σ = σ0 = 1 (dashed black line), our model Eq.(7) with σ0 = 1 (solid red line), as well as the permeability computed by Cercignani et al.36 for the Boltzmann equation (solid blue line). The Navier-Stokes equation with constant accommodation coefficients, Eq. (5), reproduces both the small h (large Kn) and the large h

(small Kn) limits, but fails to exhibit a minimum around Kn ≈ 1. Cercignani et al.’s solution to the Boltzmann equation shows the expected minimum, at the expense of an unphysical divergence at small h. This is a well-known drawback of the linearized Boltzmann equation with (geometry-independent) Maxwell boundary conditions. This divergence is enhanced in the limit of entirely specular reflection: for free molecular flow, the flow rate diverges even faster as σ0 decreases. In contrast, our Eq. (7) satisfies all requirements with a simple analytical expression adaptable to any channel geometry.

ACS Paragon Plus Environment

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 4. Permeability calculated for Eq. (7) of the present model (solid red line), for Cercignani’s solution of Boltzmann equation (solid blue line), for Arkilic model from Eq. (5) (dashed black line), and experimental data at ∆=15 psi (black dots) as a function of the

rarefaction parameter b=1/Kn.

The volumetric gas flows measured at △P varying between 1 and 15 psi or at varying nanochannel size (△P =15 psi) are represented in Figure 5A and B, respectively. Results are shown for membranes with h = 2.5, 3.5, 20, 50, 200, 250 nm (n=30 replicate measurements), where the standard deviation (always < 7%) is smaller than the size of data marks. Plotted lines are Eq. (7) with σ0 = 1. Striking agreement between the experimental data and our model (Eq. 7) was found for all values of h, which is remarkable considering that no free parameters were used.

ACS Paragon Plus Environment

Page 14 of 22

Page 15 of 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

Figure 5. (A) Volumetric gas flow Q as a function of ∆P for membranes with h = 2.5 (black squares), 3.5 (black circles), 20 (blue squares), 50 (blue circles), 200 (red squares), 250 (red circles) nm, respectively. The solid lines represent Eq. (7) for the considered values of h. (B) Volumetric nitrogen flow at △P =15 psi as a function of the nanochannel height.

Although other works considered a geometry-dependent mean free path in the boundary conditions34,35, our approach has the distinct advantage to provide a systematic way of computing the mean free path in a channel with any geometry. As pT(l) is known, one only needs to compute the path length distribution pg(l), which is only determined by wall-wall collisions and can be evaluated numerically for arbitrary geometry. The geometry-dependent effective mean free path λ0 can then be computed from Eq. (2) to any desired accuracy, and from the latter, the geometry-dependent accommodation coefficient can be obtained. Inserting the latter in the Navier-Stokes equation (5), the volumetric flow for the entire span of Kn can be quantitatively predicted, as with (7).

4. Conclusions In this paper we developed a statistical model for the calculation of the effective mean free path that takes into account particle-particle and particle-wall collisions. The nanochannel membranes employed in this work, also have unique characteristics: they achieved the high standard of accuracy and reproducibility, unmatched in the literature to date, allowing us to reliably test our statistical approach to nanoconfined gas transport. The remarkable agreement between model and experiments validated both. Such compelling outcomes is of great interest for numerous applications: in the natural gas industry our predictive model could provide a valuable tool to

ACS Paragon Plus Environment

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

improve extraction from shale formations and tight sands37, where gas is trapped in complex networks of micro- and nanopores. Refined modeling of gas flow in shale fractures could lead to increased yield, efficiency and environmental safety of gas recovery. For analytical technologies, a better description of gas transport across chromatography columns could lead to more efficient instrumentation and reduced time of analysis. Further application examples include nanofluidics for the detection of chemical agents, membranes for physical separation of gases in the chemical industry, analysis of confined fluids17,38 and controlled drug release39,40,41,42.

Acknowledgements The authors are grateful to Thomas Geninatti, Giulia Rizzo for their help on the experimental results, and to Arturas Ziemys for precious discussions. The author thank Virginia Facciotto ([email protected]) for the graphic design of schematics. Funding support from NIH NIGMS (R01 GM 127558) and CASIS (GA-2013-118, GA2014-145). Membranes provided by NMS. M.F. disclose a financial interest in NanoMedical Systems, Inc., Austin, Texas. All other authors declare no competing financial interest.

Supporting Information available including Smoluchowski equation for mean free path, probability density function from Holmin at al., membrane fabrication, and COMSOL simulation of gas flow across 250 nm membrane.

References (1)

Wu, T.; and Zhang, D. Impact of Adsorption on Gas Transport in Nanopores. Sci.

Rep. 2016, 6, 33461.

ACS Paragon Plus Environment

Page 16 of 22

Page 17 of 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

(2)

Ding, L.; Wei, Y.; Li, L.; Zhang, T.; Wang H.; Xue, J.; Ding, L.-X.; Wang, S.; Caro,

J.; Gogotsi, Y. MXene Molecular Sieving Membranes for Highly Efficient Gas Separation. Nat. Commun. 2018, 9, 155. (3)

Park, H. B.; Kamcev, J.; Robeson, L. M.; Elimelech, M.; Freeman, B. D. Maximizing

the Right Stuff: the Trade-off between Membrane Permeability and Selectivity. Science, 2017, 356, eaab0530. (4)

Sholl, D. S. and Lively, R. P. Seven Chemical Separations to Change the World.

Nature, 2016, 532, 435-437. (5)

Tong, H. D.; Jansen, H. V.; Gadgil, V. J.; Bostan, C. G.; Berenschot, E.; van Rijn, C.

J. M.; Elwenspoek, M. Silicon Nitride Nanosieve Membrane. Nano Lett. 2004, 4, 283287. (6)

Chang, N.; Gu, Z.-Y.; Wang, H.-F.; Yan, X.-P. Metal–Organic-Framework-Based

Tandem Molecular Sieves as a Dual Platform for Selective Microextraction and HighResolution Gas Chromatographic Separation of N -Alkanes in Complex Matrixes. Anal. Chem. 2011, 83, 7094-7101. (7)

Sih, J. ; Bansal, S. S.; Filipini, S. ; Ferrati,S. ; Raghuwansi, K.; Zabre, E. ; Nicolov,

E.; Fine, D.; Ferrari, M.; Palapattu, G.; Grattoni, A. Characterization of Nanochannel Delivery Membrane Systems for the Sustained Release of Resveratrol and Atorvastatin: New Perspectives on Promoting Heart Health. Anal. Bioanal. Chem. 2013, 405, 15471557. (8)

Ferrati, S.; Nicolov, E.; Zabre, E.; Geninatti, T.; Shirkey, B. A.; Hudson, L.; Hosali,

S.; Crawley, M.; Khera, M.; Palapattu, G.; Grattoni, A. The Nanochannel Delivery

ACS Paragon Plus Environment

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

System for Constant Testosterone Replacement Therapy. J. Sex. Med. 2015, 12, 13751380. (9)

Filgueira, C. S.; Nicolov, E.; Hood, R. L.; Ballerini, A.; Garcia-Huidobro, J.; Lin, J.

Z.; Fraga, D.; Webb, P.; Sabek, O. M.; Gaber, A. O.; Philips, K. J.; Grattoni, A. Sustained Zero-Order Delivery of GC-1 from a Nanochannel Membrane Device Alleviates Metabolic Syndrome. Int. J. Obes. 2016, 40, 1776-1783. (10)

Farina, M.; Chua, C.; Ballerini, A.; Thekkedath, U.; Alexander, J.; Rhudy, J.;

Torchio, G.; Fraga, D.; Pathak, R.; Villanueva, M.; Shin, C.; Niles, J.; Sesana, R.; Demarchi, D.; Sikora, A.; Acharya, G.; Gaber, A.; Nichols, J.; Grattoni, A. Transcutaneously Refillable, 3D-Printed Biopolymeric Encapsulation System for the Transplantation of Endocrine Cells. Biomaterials, 2018, 177, 125-138. (11)

Farina, M.; Alexander, J.; Thekkedath, U.; Ferrari, M.; Grattoni, A. Cell

Encapsulation: Overcoming Barriers in Cell Transplantation in Diabetes and Beyond. Adv. Drug Deliv. Rev. 2018, doi: 10.1016/j.addr.2018.04.018. (12)

Ju, Y. S.and Goodson, K. E. Phonon Scattering in Silicon Films with Thickness of

Order 100 nm. Appl. Phys. Lett. 1999, 74, 3005-3007. (13)

Knudsen, M. Die Molekularströmung der Gase durch Offnungen und die Effusion.

Ann. Phys. 1909, 333, 999-1016. (14)

Roy, S.; Raju, R.; Chuang, H. F.; Cruden, B. A.; Meyyappan, M. Modeling Gas

Flow through Microchannels and Nanopores. J. Appl. Phys. 2003, 93, 4870-4879. (15)

Ewart, T.; Perrier, P.; Graur, I. A.; MeOlans, J. G. Mass Flow Rate Measurements in

a Microchannel, from Hydrodynamic to near Free Molecular Regimes. J. Fluid Mech. 2007, 584, 337-356.

ACS Paragon Plus Environment

Page 18 of 22

Page 19 of 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

(16)

Bruno, G.; Di Trani, N.; Hood, R. L.; Zabre, E.; Filgueira, C.S.; Canavese, G.; Jain,

P.; Smith, Z.; Demarchi, D.; Hosali, S.; Pimpinelli, A.; Ferrari, M.; Grattoni, A. Unexpected Behaviors in Molecular Transport through Size-controlled Nanochannels down to the Ultra-nanoscale. Nat. Commun. 2018, 9, 1682; (17)

Grattoni, A.; Gill, J.; Zabre, E.; Fine, D.; Hussain, F.; Ferrari, M. Device for Rapid

and Agile Measurement of Diffusivity in Micro- and Nanochannels. Anal. Chem. 2011, 83, 3096-3103. (18)

Gruener, S. and Huber, P. Knudsen Diffusion in Silicon Nanochannels. Phys. Rev.

Lett. 2008, 100, 064502. (19)

Maurer, J.; Tabeling, P.; Joseph, P.; Willaime, H. Second-Order Slip Laws in

Microchannels for Helium and Nitrogen. Phys. Fluids 2003, 15, 2613-2621. (20)

Tomita, T.; Nakayama K.; Sakai, H. Gas Separation Characteristics of DDR Type

Zeolite Membrane. Microporous Mesoporous Mater. 2004, 68, 71-75. (21)

Kim, H. W.; Yoon, H. W.; Yoon, S.-M.; Yoo, B. M.; Ahn, B. K.; Cho, Y. H.; Shin,

H. J.; Yang, H.; Paik, U.; Kwon, S.; Choi, J. Y.; Park, H. B. Selective Gas Transport Through Few-Layered Graphene and Graphene Oxide Membranes. Science 2013, 342, 91-95. (22)

Cooper, S. M.; Cruden, B. A.; Meyyappan, M.; Raju, R.; Roy, S. Gas Transport

Characteristics through a Carbon Nanotubule. Nano Lett. 2004, 4, 377-381. (23)

Fine, D.; Grattoni, A.; Hosali, S.; Ziemys, A.; De Rosa, E.; Gill, J.; Medema, R.;

Hudson, L.; Kojic, M.; Milosevic, M.; Brousseau III, L.; Goodall, R.; Ferrari, M.; Liu. X. A Robust Nanofluidic Membrane with Tunable Zero-Order Release for Implantable Dose Specific Drug Delivery. Lab. Chip 2010, 10, 3074.

ACS Paragon Plus Environment

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(24)

A. Grattoni, E. De Rosa, S. Ferrati, Z. Wang, A. Gianesini, X. Liu, F. Hussain, R.

Goodall, M. Ferrari, Analysis of a Nanochanneled Membrane Structure through Convective Gas Flow. J. Micromechanics Microengineering 2009, 19, 115018. (25)

Karniadakis, G.; Beskok, A.; Gad-el-Hak, M. Micro Flows: Fundamentals and

Simulation. Appl. Mech. Rev. 2002, 55, B76. (26)

Visco, P.; van Wijland, F.; Trizac, E. Collisional Statistics of the Hard-sphere Gas.

Phys. Rev. E 2008, 77, 041117. (27)

Stops, D. W. The Mean Free Path of Gas Molecules in the Transition Régime. J.

Phys. Appl. Phys. 1970, 3, 685-696. (28)

Lorentz, H. A. Le Mouvement des Electrons dans les Métaux. Arch. Neerl. 1905, 10,

336. (29)

Bourgain, J.; Golse, F.; Wennberg, B. On the Distribution of Free Path Lengths for

the Periodic Lorentz Gas. Commun. Math. Phys. 1998, 190, 491-508. (30)

Holmin, S.; Kurlberg, P.; Mansson, D. https://arxiv.org/abs/1702.08096, 2017.

(31)

Paik, S. T. Is the Mean Free Path the Mean of a Distribution?. Am. J. Phys. 2014, 82,

602-608. (32)

Russ, S.; Zschiegner, S.; Bunde, A.; Kärger, J. Lambert Diffusion in Porous Media in

the Knudsen Regime: Equivalence of Self-diffusion and Transport Diffusion. Phys. Rev. Lett. E 2005, 72, 030101. (33)

Arlemark, E. J.; Dadzie, S. K.; Reese, J. M. An Extension to the Navier–Stokes

Equations to Incorporate Gas Molecular Collisions with Boundaries. J. Heat Transfer 2010, 132, 041006-1.

ACS Paragon Plus Environment

Page 20 of 22

Page 21 of 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

(34)

Guo, Z. L.; Shi, B. C.; Zheng, C. G. An Extended Navier-Stokes Formulation for Gas

Flows in the Knudsen Layer near a Wall. Europhys. Lett. 2007, 80, 24001. (35)

Arkilic, E. B.; Schmidt, M. A.; Breuer, K. S. TMAC Measurement in Silicon

Micromachined Channels. Rarefied Gas Dyn. 1997, 20, 6. (36)

Cercignani, C.; Lampis, M.; Lorenzani, S. Variational Approach to Gas Flows in

Microchannels. Phys. Fluids 2004, 16, 3426-3437. (37)

Javadpour, F.; Fisher, D.; Unsworth, M. Nanoscale Gas Flow in Shale Gas

Sediments. J. Can. Pet. Technol. 2007, 46, 55. (38)

Bruno, G.; Geninatti, T.; Hood, R. L.; Fine, D.; Scorrano, G.; Schmulen, J.; Hosali,

S.; Ferrari, M.; Grattoni, A. Leveraging Electrokinetics for the Active Control of Dendritic Fullerene-1 Release across a Nanochannel Membrane. Nanoscale 2015, 7, 5240-5248. (39)

Ferrati, S.; Nicolov, E.; Bansal, S.; Zabre, E.; Geninatti, T.; Ziemys, A.; Hudson, L.;

Ferrari, M.; Goodall, R.; Khera, M.; Palapattu, G.; Grattoni, A. Delivering Enhanced Testosterone Replacement Therapy through Nanochannels. Adv. Healthc. Mater. 2015, 4, 446-451. (40)

Ferrati, S.; Nicolov, E.; Bansal, S.; Hosali, S.; Landis, M.; Grattoni, A. Docetaxel/2-

Hydroxypropyl β-cyclodextrin Inclusion Complex Increases Docetaxel Solubility and Release from a Nanochannel Drug Delivery System. Curr. Drug Targets 2015, 16, 16451649. (41)

Filgueira, C.S.; Bruno, G.; Chua, C.; Ballerini, A.; Folci, M.; Gilbert, A. L.; Jain, P.;

Sastry, J.K.; Nehete, P.N.; Shelton, K.A.; Hill, L.R.; Ali, A.; Youker, K.A.; Grattoni, A.

ACS Paragon Plus Environment

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Efficacy of Sustained Delivery of GC-1 from a Nanofluidic System in a Spontaneously Obese non-Human Primate: a Aase Study. Biomed. Microdevices, 2018, 20, 49. (42)

Chua, C. Y. X.; Jain, P.; Folci, M.; Ballerini, A.; Rhudy, J.; Gilbert, A.; Singh, S.;

Bruno, G.; Filgueira, C. S.; Yee, C.; Butler, E. B.; Grattoni, A. Nanofluidic Drug-eluting Seed for Sustained Intratumoral Immunotherapy in Triple Negative Breast Cancer. Journal of Controlled Release, 2018, 285, 23-34.

ACS Paragon Plus Environment

Page 22 of 22